(eXtended Particle Swarm Optimization) enhances standard PSO by addressing two key limitations: premature convergence and poor diversity in high-dimensional or multimodal problems.
XPSO typically outperforms PSO by on benchmark functions (Rastrigin, Ackley, Rosenbrock) but adds 15–25% more computational cost per iteration. Some XPSO variants (sometimes labeled as "XPSO" based
Unlike basic PSO where particles only learn from their personal best (Pbest) and the global best (Gbest), XPSO expands this. Some XPSO variants (sometimes labeled as "XPSO" based on "expanded learning") introduce a where particles can adopt strategies from multiple exemplars, including both global and local leaders, making the optimization process more adaptive. 2. Forgetting Ability and Diversity | Feature | Standard PSO | XPSO |
XPSO algorithms introduce modifications to the velocity update equations, swarm topology, or parameter adaptation strategies to enhance search capability, convergence speed, and solution accuracy. or restart mechanisms |
| Feature | Standard PSO | XPSO | |--------|-------------|------| | | Fixed or linearly decreasing | Adaptive, chaos-based or fuzzy | | Topology | Global best (gbest) or local best (lbest) | Dynamic multi-swarm or hierarchical | | Velocity clamping | Static bounds | Self-adaptive clamping | | Diversity handling | None | Mutation, repulsion, or restart mechanisms |